J. Med. Microbiol. - Vol. 29 (1989), 195-198 0022-261 5/89/00294 1951%10.

00
01989 The Pathological Society of Great Britain and Ireland

A simple adherence test for detection of IgM antibodies

in typhoid
L. Y. ONG, T. PANG”, S. H. LIM, E. L. TAN and S. D. PUTHUCHEARY”
Department of Genetics and Cellular Biology and *Department of Medical Microbiology, University of Malaya,
59 7 00 Kuala L umpur, Malaysia

Summary. A simple adherence test to detect IgM antibodies in patients with typhoid
is described. The test utilises the IgM-“capture” approach, in which the test serum is
applied to microtitration plate wells previously coated with anti-human IgM, followed
by application of a stained Salmonella typhi antigen suspension which shows
adherence in positive cases. By this test, 58 (95%) of 61 sera from confirmed cases of
typhoid possessed IgM antibodies to the H or 0 or both antigens of S. typhi. In
patients for whom a diagnosis of typhoid was based only on a significant Widal-test
titre, 31 (41%) of 76 sera had IgM antibodies to the H or 0 or both antigens of S.
typhi.Some cross-reactivity of the IgM antibodies was detected, especially with the 0
antigens of S. paratyphi A and B. A total of 82 sera from non-typhoidal fevers
(leptospirosis, typhus, dengue fever) showed no reactivity in this test. In normal sera
there was no detectable IgM to the 0 antigen of S. typhi and only a small number
(3.9%) had low levels of IgM to the H antigen. The significance and potential
importance of this simple, sensitive, specific and economical test is discussed.

Introduction not yet exist (Edelman and Levine, 1986). There-

Typhoid remains an important public health
problem in many parts of the tropical world,
including Malaysia, which had nearly 3000 cases in
1987. Definitive diagnosis of typhoid relies upon
successful isolation of the causative organism from
patients’ specimens such as blood, stool and urine.
In addition, especially in situations where relevant
facilities for culture are not available, serological
tests such as the Widal test are often used as
additional diagnostic aids. The many limiting
factors affecting the Widal test are well-known-
e.g., the stage of illness, the normal population
titres in endemic areas, the effects of vaccination,
non-specific cross-reactions, anamnestic response,
antibiotic therapy and technical factors. These
limitationsoften lead to difficulties in the interpret-
ation of Widal test results. Furthermore, for a
meaningful result, the test requires the availability
of paired sera; the result on a single specimen is
often inconclusive. It has been stated that a rapid
immunodiagnostic test for typhoid which is sensi-
tive, specific, simple, rapid and economical does
~~ ~
fore, a simple test capable of reliably detecting IgM
antibodies to Salmonella typhi would be a useful
adjunct to existing tests, as it would presumably
indicate recent or ongoing infections.One approach
used in the detection of specific IgM antibodies is
the IgM-‘capture’ approach in which anti-human
IgM adsorbed to a solid phase is used to bind
selectively IgM present in examined sera and thus
eliminate competition by other antibody classes in
the subsequent testing (Ryan and Kwasnik, 1985).
We report here the development of such a test-a
simple, specific and economical method to detect
IgM antibodies in cases of typhoid, using the IgM-
“capture” approach and involving adherence of
stained S. typhi antigen suspensions in the final
stage.
Materials and methods
Sera
Four groups of serum samples were collected from
different sets of individuals.
Group I consisted of 61 single samples of serum from

Received 5 Sep. 1988; revised version accepted 12 Jan. 1989.

*Correspondence should be sent to Dr T. Pang, Department of patients with a clinical diagnosis of typhoid admitted to
Medical Microbiology, University of Malaya, 59 100 Kuala the University Hospital, Kuala Lumpur. These sera had
Lumpur, Malaysia. Widal-test titres of 2640 (for both H and 0 antigens).

195
196 L. Y. ONG ET AL.

The diagnosisof typhoid in these patients was confirmed by formation of an adherence pattern, whereasa “button”
by the isolation of S. typhi from blood or faeces, or both. indicated the absence of IgM antibody.
The date of onset of clinical symptoms varied from 3-4
days to 6-7 weeks. IgM abolition by 2-mercaptoethanol treatment
Group 2 consisted of 76 single samples of serum from
patients in whom the diagnosis of typhoid was based on For removal of IgM from selected sera, 2Opl of 2-
a significant titre in the Widal test (2640 to H or 0 mercaptoethanol(2-ME ; Sigma Chemical Co., St Louis,
antigen, or both) but without confirmation by isolation MO, USA) was mixed with an equal volume of the serum
of S. typhi. and the mixture was incubated for 1 h at 37°C and then
Group 3 consisted of 82 single samples of serum from tested for IgM as described above.
patients with non-typhoidal fevers common in the study
region-leptospirosis (20), typhus (42) and dengue fever IgM removal by immunobeads
(20). Leptospirosis was diagnosed by the microscopic
agglutinationtest (Alexander, 1976), typhus by the Weil- Samples(100 pl) from selected sera were pre-incubated
Felix reaction to Proteus OXK, OX19 and OX2 antigens with an equal volume of anti-human IgM-coated im-
(Felix, 1944) and dengue fever by virus isolation (Lam et munobeads (BioRad Laboratories, Richmond, CA,
al., 1986) or haemagglutination inhibition (Clarke and USA), at a concentration of 1 mg/ml in borate-saline
Casals, 1958). buffer, at 4°C for 1 h. Sera were then tested in the usual
Group 4 consisted of sera from 102 healthy individuals manner.
from the medical-student population of the University of
Malaya. Results
Of the 61 Group-1 sera from confirmed cases of
Antigens typhoid, as indicated by isolation of S. typhi, 58
Antigen suspensions comprised stained, killed H and (95%) had IgM antibodies to the H or 0, or both,
0 antigens from S. typhi, S . paratyphi A, S. paratyphi B antigens of S. typhi by our test (table I). Within this
and Proteus OXK, OX19 and OX2 (Wellcome Reagents group, 51 (84%) had IgM antibodies to both H and
Ltd, Kent). 0 antigens, six (10%) to only the H antigen and one
(2%) to only the 0 antigen. Of the 58 IgM-positive
sera in this group, more than 80% had IgM titres of
Widal-test 21280 (table 11). In a random sample of ten of
The Widal test on various sera was performed by
standard procedures, as described previously (Pang and
Puthucheary, 1983). Table I. Levels of IgM antibody to S. typhi in various
groups of sera

Adherence test to detect IgM Number of sera Number IgM

The adherence test to detect IgM to S . typhi was Group Category in group positive*
performed by adding 100 pl of rabbit anti-human IgM
(p chain-specific,Dako Immunochemicals,Copenhagen, 1 Typhoid (confirmed) 61 58 (95%)
Denmark) at a dilution of 400 in borate-saline buffer (pH 2 Typhoid (serology only) 76 31 (41%)
9.0) to wells of U-bottom microtitration plates (Immulon 3 Other fevers 82 0
4 Normal 102 4 (3.9%)
2, Dynatech Laboratories, Alexandria, VA, USA). The
dilution of 400 was decided after testing the antisera at
various dilutions from 100 to 1600. After overnight * Positive to S.typhiH or 0 antigens, or both, at a titre of 2 80.
incubation at 4”C, coated plates were washed three times
with distilled water containing Tween 20 (0.05%) and Table 11. Titres of IgM antibodies to S. typhi H and 0
tapped dry. Plates prepared in this manner can be stored antigens in the 58 positive Group-1 sera by the adherence
at 4°C for up to 4 months. Patients’ sera were then serially test
diluted in the anti-human IgM-coated plates in normal

I
saline, starting with an initial dilution of 80 (100 pl final
volume/well). Plates were then incubated for 2 h at 37”C, Number of sera reacting to
washed three times with distilled water and tapped dry.
Then 50 1 of antigen suspension (diluted approximately
1 in 16 in normal saline) was added to the wells and the
Titre range I antigen 0 antigen

plates were incubated further at 37°C for 4 h (for 0 80-640

antigen) and overnight (for H antigen). Plates were then 1280-5 120
read by means of a microtitre mirror (Cooke Engineering 2 10 240
Co., Alexandria, VA, USA). A positive result was shown
 IGM ANTIBODY DETECTION IN TYPHOID
these IgM-positive sera-nine with titres of 1280-
B 10 240 (0 and H antigens) and one of 320 (0),
640 H-treatment for removal of IgM by 2-ME
resulted in marked falls in the 0 and H titres,
measured by our method, to undetectable levels,
except for one serum, which nevertheless showed a
very large reduction, from 2 10 240 to 640 (0),320
(H). Similar results were obtained by treatment of
the same set of sera with anti-human IgM-coated
immunobeads, as used in earlier dengue virus IgM
studies (Gunasegaran et al., 1986).These 10 sera all
gave negative results in the latex agglutination test
for rheumatoid factors (Winchester, 1976). Of 42
sera from Group 1 showing IgM to S. typhi H
antigen, two (5%) cross-reacted with the H antigen
of S. paratyphi A and one (2%) cross-reacted with
the H antigen of S. paratyphi B. All three cross-
reacting sera had higher IgM titres to S. typhi than
to S. paratyphi A or B. Of 39 Group-1 sera showing
reactivity to S . typhi 0 antigens, 13 (33%) showed
various degrees of cross-reactivity to the 0 antigens
of S. paratyphi A and B and, of these, five showed a
high degree of cross-reactivity (titres 2 10 240) to
the 0 antigen of S. paratyphi A or B .
In Group 2, for which a diagnosis of typhoid was
based solely on a significant Widal-test titre, 31
(41%) out of 76 sera tested had IgM antibodies to
H or 0, or both, antigens of S . typhi (table I). In
Group 3, none of the sera from patients with
leptospirosis (20), or typhus (42) or dengue fever
(20) reacted with S. typhi in this test (table I).
Similarly,none of the 102normal sera tested (Group
4) showed detectable levels of IgM antibody to the
0 antigen of S . typhi and only four (3.9%) had a
measurable titre against the H antigen (table I).
Discussion
An important approach to the rapid diagnosis of
infectious diseases is the detection of specific IgM
antibodies, the presence of which is usually indica-
tive of a recent or ongoing infection (Ryan and
Kwasnik, 1985). Other studies have shown a
significant increase in IgM levels (as measured by
single radial diffusion) in patients with typhoid, at
all stages of the illness, from the first week onwards
(Kumar et al., 1974). The IgM-adherence method
that we have described provides a simple, sensitive,
specific and economical test to detect IgM class
antibodies in typhoid. The test is economical in
using readily available and inexpensive materials
and reagents and in its lack of dependence on
sophisticated instrumentation. The test was able to
detect IgM in 95% of confirmed typhoid cases and
no false-positives were detected in sera from other,
197
non-typhoidal fevers common in the region, uiz.
typhus, leptospirosis and dengue fever. In compar-
ison, other studies based on an ELISA approach
detected IgM to the LPS antigen of S. typhi in
8 1*8%(Srivastava and Srivastava, 1986)and 93.1%
(Nardiello et al., 1984) of proven cases of typhoid.
In retrospect, it would have been useful to have
included the ELISA approach in the present study
for the purpose of comparison and this should
probably be considered in any future similar
investigations. Among the three typhoid patients
whose sera showed no detectable IgM in the present
study, one had been treated with chloramphenicol
and other was admitted 2 months after the onset of
symptoms, by which time the IgM levels may have
dropped to an undetectable level. In relation to this,
it is known that early chemotherapy can suppress
antibody formation in typhoid cases (Watson, 1957)
and that IgM levels may have subsided after 45-90
days (Nardiello et al., 1984). If these two cases are
excluded from the group, 58 of the remaining 59
sera (98%) had detectable IgM by this test. As
expected, the present study showed cross-reactivity
of the IgM antibodies, especially with the 0
antigens of S. typhi-related organisms (S.paratyphi
A and B ) due to the sharing of antigenic component
12 (Christie, 1980).
The present study also suggests that the IgM-
adherence test is likely to be helpful in interpreting
a single, high-titre Widal test result when isolation
of the causative organism is unsuccessful or not
attempted. The presence of IgM antibodies in 41%
of the sera in this group would provide further
diagnostic support for a diagnosis of typhoid in
such cases. However, explanations are needed for
the remaining, IgM-negative sera in this category.
It could be questioned whether they are real cases
of typhoid, but if so, it is possible that the IgM level
in these sera may have been below the sensitivity
limit of the IgM-adherence test. This seems un-
likely, in view of the fact that IgM was detectable
in 9598% of confirmed typhoid cases and that the
patients’ characteristics for the two groups (i.e.,
age, date of onset, etc.) were similar. The more
likely possibility is that these patients were suffering
from other febrile illnesses and that the elevated
Widal-test titres were thus a result of an anamnestic
response. Wicks et al. (1974) have suggested that
high Widal titres at the early stages of typhoid in
an endemic area represented an anamnestic re-
sponse. It has been suggested that anti-H antibody,
in particular, is anamnestically responsive and
easily stimulated by non-specific stimuli (Christie,
1980). Our results with these sera, which showed
high titres in the Widal test but which were negative
198 L. Y . ONG ET AL.
by isolation,also caution against making a diagnosis
of typhoid based on a single, elevated titre in the
Widal test.
It seems likely that our test for detecting IgM to
S. typhi should prove an important additional
method in the laboratory diagnosis of typhoid. In
relation to other methods previously used for this
purpose, such as ELISA (Nardiello et al., 1984;
Srivastava and Srivastava, 1986), crossed-immu-
noelectrophoresis (Tsang and Chau, 1981) and
radio-immunoassay (Tsang et al., 198 l), the IgM-
adherence test is superior in being economical and
simple to perform and in utilising non-hazardous
materials and not requiring sophisticated instru-
ments. Furthermore, it has the important advantage
of being applicable even when only a single serum
specimen is available. In our experience, about 70-
80% of sera submitted for typhoid diagnosis are
single specimens. Furthermore, in endemic areas
such as Malaysia, antibody titres may rise very
rapidly because of previous exposure and therefore
demonstration of rising titres between two speci-
REFERENCES
Alexander A D 1976 Serological diagnosis of leptospirosis. In :
Rose N R, Friedman H (eds) Manual of clinical immunol-
ogy. American Society for Microbiology, Washington, DC,
pp 352-356.
Christie A B 1980 Infectious diseases, 3rd edn. Churchill
Livingstone, Edinburgh, pp 47-102.
Clarke D H, Casals J 1958 Techniques for hemagglutination
and hemagglutination-inhibition with arthropod-borne vi-
ruses. American Journal of Tropical Medicine and Hygiene 7 :
561-573.
Edelman R, Levine M M 1986 Summary of an international
workshop on typhoid fever. Reviews of Infectious Diseases
8: 329-349.
Felix A 1944 Technique and interpretation of the Weil-Felix
test in typhus fever. Transactions of the Royal Society of
Tropical Medicine and Hygiene 37 : 32 1-341 .
Gunasegaran K, Lim T W, Ahmed A, Aaskov J G, Lam S K,
Pang T 1986 Hemadsorption immunosorbent technique for
the detection of dengue immunoglobulin M antibody.
Journal of Clinical Microbiology 23 : 170-1 74.
Kumar R, Malaviya A N, Murthy R G S, Venkataraman M,
Mohapatra L N 1974 Immunological study of typhoid:
immunoglobulins,C3, antibodies, and leukocyte migration
inhibition in patients with typhoid fever and TAB-
vaccinated individuals. Infection and Immunity 10: 1219-
1225.
Lam S K,Chew C B, Poon G K, Ramalingam S, Seow S C,
Pang T 1986 Isolation of dengue viruses by intracerebral
inoculation of mosquito larvae. Journal of Virological
Methods 14: 133-140.
Nardiello S,Pizzella T, R u m M, Galanti B 1984Serodiagnosis
of typhoid fever by enzyme-linked immunosorbent assay
mens may be difficult (Wicks et al., 1974). Thus, in
the present study, high titres were already detecta-
ble in some single sera only 3-4 days after the onset
of symptoms. IgM titres were measured only to the
H and 0 antigens because background Vi agglutin-
ins are known to be present in the normal
populations of endemic areas (Report, 1961).
Therefore, measurement of these antibodies is of
limited value both in the diagnosis of typhoid and
in determination of the carrier state. Within the
diagnosticlaboratory, the IgM-adherencetest could
be used directly as a sole serological test or,
alternatively, as a confirmatory test following
screening of sera by the Widal test. The latter
approach would be particularly useful in smaller
laboratories, for example those attached to district
hospitals, where bacteriological isolation and iden-
tification facilities may not be available.
We thank Dr V. How, Department of Medical Microbiology,
University of Malaya for providing normal sera. This work was
supported by China Medical Board Grant No. 78-374.
determination of anti-Salmonella typhi lipopolysaccharide
antibodies. Journal of Clinical Microbiology 20: 7 18-72 1.
Pang T, Puthucheary S D 1983 Significance and value of the
Widal test in the diagnosis of typhoid fever in an endemic
area. Journal of Clinical Pathology 36: 471-475.
Report of the PHLS Working Party on the Bacteriological
Examination of Waterworks Employees 1961. Journal of
Hygiene 59 : 231-247.
Ryan R W, Kwasnik I 1985 Specific immunoglobulin detection.
In :Lennette E H et al. (eds) Manual of clinical microbiology,
4th edn. American Society for Microbiology, Washington,
DC, pp 877-882.
Srivastava L, Srivastava V K 1986 Serological diagnosis of
typhoid fever by enzyme-linked immunosorbent assay
(ELISA). Annals of Tropical Paediatrics 6 : 191-194.
Tsang R S W, Chau P Y 1981 Serological diagnosis of typhoid
fever by counterimmunoelectrophoresis. British Medical
Journal 282 : 1505-1 507.
Tsang R S W, Chau P Y,Lam S K, La Brooy J T, Rowley D
1981 Antibody response to the lipopolysaccharide and
protein antigens of Salmonella typhiduring typhoid infection
I. Measurement of serum antibodies by radioimmunoassay.
Clinical and Experimental Immunology 46 : 508-5 14.
Watson K C 1957 The relapse state in typhoid fever treated with
chloramphenicol. American Journal of Tropical Medicine
and Hygiene 6 : 72-80.
Wicks A C B, Cruickshank J G, Musewe N 1974 Observations
on the diagnosis of typhoid fever in an endemic area. South
African Medical Journal 48: 1368-1370.
Winchester R 1976Tests for detection of the rheumatoid factors.
In: Rose N R, Friedman H (eds) Manual of clinical
immunology. American Society for Microbiology, Wash-
ington, DC, pp 665-668.